Helm Minutes for 2018

December 20, 2018
Announcements
No meeting next week!
Helm 3 repos working group, Wednesdays 8:30-9:00 am PST [link]
Helm Hub is live! https://hub.helm.sh
Helm 2.12.1 is out (decaf edition)
Standup (kubecon was last week)
Adam
Giving update on Kubecon/Helm v3
The v3 talk was highly attended and engaged


Matt Fisher
Josh
Matt Farina
Michelle
Scott
Adnan
Helm v3 Status
https://github.com/helm/helm/projects/1
Tasks open for people to take on are tagged with “help wanted”
In the past week we organized to update the details on the tasks
Discussion
Proposal for CRD Support
We would like less intrusive changes to support CRDs
Will CRDs become GA or overcome as TPRs were?
Will look into patterns in current system that we could potentially use
Helm 2 to 3 upgrade story
A tool that can migrate from v2 config, as managed by tiller (configmaps/secrets), to the new v3 format
Document what we need to do (proposal in community repo)
Helm 3 repos update (Josh)
TLDR; Helm charts will be able to be stored in an image registry
https://groups.google.com/a/opencontainers.org/forum/#!topic/dev/qYM84b0vays
Assignments
Issue Sherpa: Michelle
Notes: Josh
Moderator: Matt Farina
December 6, 2018
Announcements
Helm Deep Dive now 80 minutes (2 back to back time slots) at CloudNativeCon
First half is the Helm 3 deep dive, second half is free-form discussion
Helm 2.12.0-rc.2 has been released
Standup
Matt Fisher
Working on the helm 2.12 release. Went through the PR queue to see if there are any fixes. RC2 has been release. Plan is to cut release over the next couple of days before KubeCon.
Attending to the helm repository discussion
Matt Farina
Working on the hub to get launched
Taking the charts in image repos to the OCI
https://groups.google.com/a/opencontainers.org/forum/#!topic/dev/U49AZRFo3tk
Nikhil
Back from paternity leave - just getting back into the swing of things.
Josh
Working on chartmuseum 0.8 with bearer auth support
Continuing to moderate the helm repository discussion
Now discussing with people at docker on the distribution API
Brian
Working on utilities on mapping Go values to Lua, deserializing back to YAML
Adnan
Working on the hub
Post-kubecon, looking to add some CI to deploy to the helm hub repo
Continuing development on monocular
Justin
Busy, hoping to help out with the issue queue/PR queue
Taylor
Helping fisher on the 2.12 release
Want to have a discussion on adding contributors to help out on Helm 3 (onboarding)
Michelle
Putting together some slides on Helm 3, hub.helm.sh and Helm 2 for the KubeCon keynote update from Liz
Helm v3 Status
Adam has no voice/internet today, therefore no update this week
Discussion:
Taylor: on adding contributors to help out on Helm 3 (onboarding)
Taylor: action item to talk to adam about discuss onboarding
Some bugs that may need some love: 949, 990, 1098, 1168, 1193, 1237, 1354
Discussion: how do we want to handle long-standing issues that have gone stale?
Farina: we should probably come up with a process for closing out issues to close an issue when it hasn’t been implemented in a given period of time from the community or the maintainers
Action item: michelle will open a ticket for discussing the issue of closing long-standing feature requests
How do we get proposals in?
This process should be documented
Action item: henry will open the issue to write up the docs
Should we have another standup next week?
Yes, one standup next week then cancel until the new year
Assignments (for week after Kubecon - 20th )
Issue sherpa: matt, matt and michelle all working part-time
Notes: (Nikhil can’t make this) Farina
Moderator: taylor

November 29, 2018
Announcements
Recurring Helm 3 Repos discussion, Wednesdays 8am PST
Standup
Josh
Setup weekly Helm 3 Repos discussion, Wednesdays 8am PST
Had call #1, made some progress.
Removing --version flag
Keeping “helm repo add”
Matt Farina
Worked on Helm Intro slides for seattle
Signed up for the project maintainers booth in seattle
Matt Fisher
Adam


Helm v3 Status
Helm repo stuff, not ready for Kubecon
Requirements handling, cleanup. Merging arbitrary data. Pre-req for adding Lua to execution engine


Discussion
Fisher: Helm 2.12 release status
1 PR - feature flag to helm upgrade “--cleanup-on-fail”
https://github.com/helm/helm/pull/4871
Needs few more reviews (Taylor + Fisher + ?)
Can this make it in? https://github.com/helm/helm/pull/4709 (Jacob Legrone)
Butcher: looks great
Is there a plan for a helm contributors/maintainers social get-together, drinks (I volunteer to buy the first round!) at kubecon? [henrynash]
CNCF booth? Microsoft Booth?
Look for peoples hair
MS Reactor event: https://open.microsoft.com/2018/11/21/microsoft-pre-kubecon-community-workshops/


Fixing bugs in v2 & v3 [mattfarina]
For features - if they do not affect Helm 3, we consider merging
If it solves for Helm 2 and 3, asking contributors if they are willing to add these things for Helm 3
Bugs - will be merged
System wide plugins [mattfarina]
SUSE wants to add this for envs that are unable to install plugins
Can be handled by https://standards.freedesktop.org/basedir-spec/basedir-spec-latest.html
Adam: this is on Helm 3 project board , XDG configuration
Given the increasing agenda length, if we are going to stick to a 30min call, maybe we should skip the Standup section? Alternatively do we need an hour call?  [henrynash]
Fisher: there was discussion on the cncf-helm-cartographers mailing list about everyone writing out their standup and dedicating the 30 minutes to discussion. If the standup topic requires input, it can be added to discussion items.
Standup is good to give people chance to speak up, get people engaged
Stef: comes to meeting to keep up with what were up to, also to be involved
Technical-level discussions, fostering this type of community at SUSE
Henry:
Martin: just make standup shorter, main points in discussion
Assignments
Moderator: Josh
Notes: Matt Fisher
Issue Triage: Martin
^ may not have access to add labels etc

November 22, 2018
Cancelled due to holidays

November 15, 2018
```yaml
Announcements
Brian is the bee’s knees: github.com/Azure/golua
Because of Thanksgiving, this meeting will be skipped on Nov 22.
Standup
Josh
Meeting on tuesday about the helm repo discussion, most people on the call are in favour as long as we can ensure that we have the resources to commit to that work
Will send out an email to the CNCF mailing list
Matt Fisher
Been working on getting the helm 2.12 release out
Michelle
Trying to figure out two issues at the moment: the helm linter not linting subcharts as well as openshift SecurityContextConstraints (would love assistance on the latter)
Brian
Been working on golua as well as some debug facilities into the VM execution for the lua engine
Spending the next few weeks focusing on the debugging library as well as panic traces in the lua VM
Matt Butcher
On the Helm 3 side, opening PRs on re-naming some of the commands
Working on the Helm 3 conference stuff for KubeCon Seattle
Attempting to make progress on a Helm 3 alpha release for KubeCon Seattle (tentative)
2 sessions for the core maintainers: intro and advanced sessions
Working on a Helm Summit EU for spring 2019
Adnan
Focusing on Monocular and kubeApps
(butcher) are we planning on a KubeCon announcement for that one?
(adnan) yeah, it’d be great to announce that in the intro
(butcher) it might be getting a main stage annoucement
Matt Farina
We might be getting the main stage for KubeCon
Working on a session for the Hub
Issues triaging etc.
Adam Reese
Looking into integrating golua into Helm
Mostly focusing on helm 3
Helm v3 Status
Discussion
Fixing bugs in v2 & v3 [mattfarina]
What is the success criteria for going back to one issue triager? [bacongobbler]
Will follow up in cncf-core-maintainers mailing list
At what point can we reasonable close an issue for being stale/having no activity? From last week’s convo, we decided not to have a bot automatically close issues, but can we decide on a time limit that a triager can use. 6 months? 8 months? 1 year? [mnoorali]
System wide plugins [mattfarina]
https://github.com/helm/helm/pull/4871 [mattfarina]
CRD lifecycle - Helm and deleting CRDs (doesn’t today) and CRD lifecycle [mattfarina]
The crd-install hook only deals with the install lifecycle but there are open issues IRT CRDs with Helm. How should we approach this?
Adam: we don’t do lifecycle management around CRDs because CRDs can be shared across multiple charts, so if we delete a CRD on a `helm delete`, then we could delete a CRD that could be dependent on other charts
James: I agree, but with the introduction of the annotation, we’ve seen weird behaviour where if someone has created a CRD prior to a `helm install`, the CRD will get removed before being installed. We’re trying to figure out what we can do to make the situation better for istio/cert-controller
Adam: I haven’t dug too deep into the CRD APIs, but perhaps an action item we can do is observe the behaviour how the Kubernetes API handles CRDs outside of Helm, and determine the expected behaviour? Perhaps even a few concrete use cases
Mortent: would it be valuable to add a proposal for the solution in Helm 3?
James and Adam: yeah, this might affect Helm 3 as well so if there are fixes that need to happen there we should get that in
At IBM we have several hundred charts, and I hear this issue of CRD quite a lot - I’ll do a review and see if I can bring requirements to the table [henrynash]
<heresy>Given the increasing agenda length, if we are going to stick to a 30min call, maybe we should skip the Standup section?</heresy> [henrynash]
Adam: should we talk about this in 2 weeks at next standup?
Henry - happy with that
Given we are entering a period of large new features for helm, would it be beneficial to formally articulate/document how new features should be agreed for inclusion? Many open source projects have a spec repo (e.g.: https://github.com/openstack/keystone-specs), where the PR for a spec getting merged means that the feature and its architectural implementation is agreed. I guess our equivalent is the helm/community/proposals|helm-v3 ? However, it is not clear how such things really get accepted, what’s the threshold for sign off by maintainers etc. [henrynash] (let’s defer this to 2 weeks time)
PR #4871? [farina & fisher]
Adam: Let’s follow up offline
Assignments
Moderator: Fisher
Notes: Josh
Issue: Michelle & Fisher


November 8, 2018
Announcements
Because of Thanksgiving, this meeting will be skipped on Nov 22.
Standup
Matt Farina
Specs repo, document how everything works
Issue sherpa-ing this week, trying to bring down queue
Changes to monocular to work w MongoDB over SSL, to support Helm Hub
Matt Butcher
SSL issue, CRLs not supported by GRPC/Go crypto libs (link?)
Going away in Helm 3, so if somebody else wants to take it, go for it
Formal specification for Helm 3 charts - to track what is and isnt supported
Some parts of charts in Helm 2 are unclear/not documented
Starting on Helm 3 CLI stuff soon
Taylor
Slide decks for Kubecon shanghai
Ask taylor for help looking at PRs
Michelle
Reviewed old issues
Parallel flag on Helm test, did testing (link?)
More issue triage this week
Josh
Looking into a proposal that Steve put together involving repositories for v3 -> #55
Submitted another related proposal. Will talk about that in discussion. -> #57
Matt Fisher
Issue triaging this week
Primarily focused on PR queue
Slating things for 2.12 RC release
Adam
Helm 3 CRDs
Some PR reviews
Adnan
Working on Monocular w Farina
Submitted PR to monocular to support
Monocular chart still needs to allow Mongo params as secret vs. values
Brian
Stabilizing Lua VM
Prototyping events code
Helm v3 Status
Project board: https://github.com/helm/helm/projects/1
Cards are being added linked to topics from the proposal in helm/community
Will be converted to issues when people are ready
Lua, charts format in progress
Kubernetes-core removed as a dependency
Discussion
Helm v3 repository topics [jdolitsky]
Community proposal #55 - merging helm repos and docker registries
The basis of this proposal is for helm repositories to mimic the docker registry API more closely
Looking on what is the right way to build it out further
Jimmy: the problem I see is the backwards compatibility for the pre-existing repository system (referring to Helm 2 repositories)
Fisher: the other issue comes in the form of OCI mediaType incompatibility (you can’t helm push quay.io/bacongobbler/helloworld and docker push quay.io/bacongobbler/helloworld)
The quay implementation restricts repositories to a single mimetype to avoid this problem, but that has UX tradeoffs
Follow-up: let’s set up a call to discuss this meeting (assigned to Josh Dolitsky)
Community proposal #57 - Helm 3: push, login, capabilities api
Assignments
Moderator: Adam
Notes: ?
Issue:
Michelle
November 1, 2018
Announcements
Josh and Scott are now Helm Org Maintainers [mattfarina]
Standup
Josh
Working on chartmuseum 0.8 with token bearer auth support
Worked on proposal with Azure folks on merging docker and helm repos
Adam
Last week in a company team offsite
Matt Fisher
In team offsite last week
Doing triaging for v2.11.1 looking to see if we should patch release
Michelle
Responded to comments on helm test proposal (for v3)
https://github.com/helm/community/pull/40
In team offsite last week
Matt Butcher
Team meetings last week
Working on helm v3 command line argument changes
Removed moniker naming
Brian
At team offsite
Focused on Helm v3 lua integration
Matt Farina
Doing issue triaging last week
Working on nested repositories for the Helm 3 proposals
Taylor
Writing slides for China and Helm deep dive
Helm v3 Status
Talked through the board
When something in the proposal goes to being built we can:
Convert existing cards to issues in the project
When implementation happens we should add a link to an implementation issue where further details can be found in the proposal
Discussion
How can we handle the 500+ open issues? [mattfarina]
Not going to add a stale issue closer yet but discussed
Going to have multiple issue sherpas to bring the number down
Issue #3943 - allow a fix for this in helm 2? [jdolitsky]
This fix can go in as long as it’s opt-in and no APIs are changed in backward incompatible ways (including tiller api)
Assignments
Moderator: Matt Butcher
Notes: Josh Dolitsky
Issue:
Matt Farina
Matt Fisher
October 25, 2018
Announcements
Distributed charts repo search update via mailing list email
Standup
Josh
Investigating helm login commands for different repository providers
ChartMuseum has a PR for oracle storage backend
Matt Farina
Worked on some charts tooling
Showcased helm repo audit
Taylor
Not much coding this week
Prepping to talk in China about v3
Discussion
Add weekly review of helm 3 github project to this meeting [Adam Reese]
Issues needing some extra love:
Potentially can try and close out: 4789, 4773
Need a response to: 4833, 4831, 4827, 4826, 4824, 4784, 4767, 4754, 4744, 4699, 4695, 4679, 4698, 4697, 4695, 4679
Assignments
Moderator: Josh
Notes: Matt Fisher
Issue: Matt Farina

October 18, 2018
Announcements
No announcements for this week
Standup
Matt Farina
Working on bringing back the PR sizing stuff
Looking back on the hub stuff with adnan in the future
Matt Fisher
Not much to report this week
Taylor
Working on a talk for KubeCon Shanghai on an intro to Helm
Also looking at doing a talk on a Helm 3 deep dive, also at KubeCon Shanghai
Adam
Started putting together a github project for Helm 3 work
https://github.com/helm/helm/projects/1
Non-core maintainer standup
Martin Hickey
Dropped a few issues down in chat
Issues potentially can try and close out: 4782, 4775, 4768, 4751, 4745, 4713, 4705, 4689
Issues need a response to (potential quick response): 4796, 4793, 4790, 4767, 4744, 4737, 4729, 4712, 4699, 4698, 4697, 4695, 4679
Discussion
edX course… anyone interested in doing it? CNCF has $$ [Matt Farina]
University-like material for education on Helm
Taylor: I might be interested in it but not sure about the timeframe is on that
Farina: let’s talk about this offline, but this opportunity just came up
Should helm login support basic auth in addition to oauth? [Matt Farina]
Adam: I’m not sure how difficult it’d be to add that in
Farina: so we should probably support an oauth mechanism out of the box for now
Helm container image? What base image (scratch, alpine, etc?) [Matt Farina]
Tiller uses alpine, so we should probably re-use that
Follow-up question: should we also push tiller images to quay/dockerhub?
I think we’re at an acceptable time to push to both gcr and quay using CircleCI
Linkerd2 has a different way to inject sidecars. How can we do this with Helm today? [Matt Farina]
Adam: yeah, I don’t think we could accommodate this today, essentially linkerd2 needs to stream-edit the chart as it’s being applied
Farina: could the event changes in Helm 3 help accommodate this?
Adam: theoretically, yes
Farina: perhaps it would be good to sit down with the linkerd folks
Assignments
Moderator: Taylor
Notes: Farina
Issue: Fisher
October 11, 2018
Announcements
Matt Butcher is the new chair of the org maintainers
Standup
Josh
Looking into v2 repo index
Challenges are that it is partitioned into multiple files
Design work around things like nested chart directories needed
Farina
Working on getting the Hub site rolling
Dco-labeler moved to charts AKS cluster
Issues with AKS cluster
Adnan
Helm hub site
Monocular 1.0
Cosmos DB
Fisher
Issue queue stuff
Adam
Have some PRs to go through, small changes, docs changes
Out for a bit
Focusing on helm 3 stuff tomorrow
Kubernetes 1.12 merged
Michelle
Went through some PRs this morning
3805 and 1193
Brian
Working on lua for Helm 3, continuing through the tests
Notes and talking points for the use of Lua in Helm
Justin
A little issue triage
Hoping to jump on some PRs soon
Butcher
Helm trademark move from Deis/MS to CNCF
KubeCon Helm stuff
Planning work on Helm Summit in the EU (spring next year)
Discussion
How can we better communicate around Helm v3 and let people help with it? [Matt Farina]
[Henry Nash] Maybe at least add weekly item for an update on v3 development, calls for help, focus areas etc. ?
Implementing index v2

Assignments
Moderator: Adam
Notes: Fisher
Issue: Michelle
October 4, 2018
Announcements
Helm Org Maintainers vote link [Matt Butcher and Matt Farina]
Standup
Adnan
Working on Monocular 1.0, migrating onto CircleCI because that’s what most teams are using it seems
Josh
In talks with the Azure Container Registry team in regards to API design
Matt Farina
Working on the hub site (hub.helm.sh)
All the infra is there and ready (cert-manager, letsencrypt, ingress etc.)
How do we store secret information? Want to chat about it today
Matt Butcher
Transferring the Helm trademark to the CNCF, working with the Microsoft lawyers for transferring the paperwork in time for KubeCon
Working on the Helm 3 stuff with Adam and Brian
Brian: working his way through implementing the Lua stdlib
Vic
I have been getting back into reviews on the charts repo.
Matt Fisher
Last week has been maintaining the issue queue, ensuring whether or not we need to do a v2.11.1 release
Adam
Last week has been focused on Helm 2 and Kubernetes 1.12 support
Nothing ready to show yet, waiting on upstream to fix some things which should be coming out later today
Non-core maintainers
Martin Hickey
Is it true that the Helm 3 codebase has been locked down to core maintainers?
(adam) yeah, there’s a ton of refactoring going on in the core codebase, but there are a couple areas where help would be appreciated (documentation, testing infra possibly?)
In terms of docs, what areas can we help?
(adam) I’m currently looking at sharing some documentation around what’s changed, possible opening a github project to show what’s being worked on/what’s done/what’s in the backlog
Markdown docs are stale in dev-v3, so updating the README or jumping into existing project work (docs site, CI infra, repository API working group)
Discussion
Encrypting secrets for sites and services [Matt Farina]
Org maintainers - congrats! You are now responsible for the following [Josh]:
Maintaining the mission, vision, values, and scope of the project
What is it that Helm trying to solve? (mission)
Where does Helm start and end? (scope)
How will we get there? (vision)
What do we need to embrace in order to succeed? (values)
Refining the governance and charter as needed
What can we do to make this process ultra-efficient?
How do move fast and not break things?
Making project level decisions
How can we make sure people have necessary info in order to confidently vote?
Which type of decisions are made at the subproject level? (eg. choice of CI, language, etc.)
Resolving escalated project decisions when the subteam responsible is blocked
What are the respective subteams?
Who will be primary points of contact for each?
Managing the Helm brand
Websites, logos
Documentation
Tutorials and evangelism
Twitter, social media, etc.
?
Controlling access to Helm assets such as source repositories, hosting, project calendars
Which tools will we use managing and sharing credentials?
Handling code of conduct violations
Violators will be publicly executed with a 2/3 vote
Deciding what sub-groups are part of the Helm project
Should this divide into more than just subprojects? (eg. testing committee)
Overseeing the resolution and disclosure of security issues
Majority of issues would related to Kubernetes in Helm 3
Does CNCF have a process in place for this?
Managing financial decisions related to the project
What types of things might cost money?
How does being a CNCF project relate to this?
Can companies or individuals contribute?
How can we ensure that the direction of the project cannot be “bought”?
Assignments
Moderator: Josh
Notes: Adnan
Issue: Adam
September 27, 2018
Announcements
Helm 2.11 with Kubernetes 1.11 support was released
Tomorrow is the last day for nominating Helm project maintainers
Standup
Adam
On vacation last week
Continued work on Helm 3
Started on Kubernetes 1.12 support
Farina
Working on getting the Hub cluster up and going
Published a blog post on how to use chart-testing yourself!
Moving bots under the helm org
Butcher
Provenance and crypto stuff
Blog post for the Helm blog
Staying on top of Helm 3 things
Brian
Get the lua implementation passing against the spec conformance test suite
Issues around snapshotting values, ironed them out
Continuing to wrap up lua implementation over the next few weeks
Justin
Issue triage, helped people close a few issues
Adnan
Monocular 1.0 close to cutting
Working on moving the chart-repo and chartsvc to the Helm org, moving images to quay
Fisher
Going through Helm 2.11 release
Ensuring RCs have been going through smoothly
Stef
SUSE working on a Helm mirror utility
Discussion
KubeCon Seattle Helm Talk Coordination [Matt Butcher]
3 talks: intro to Helm, draft, deep dive on Helm
Reddit is going to talk about their story for getting into k8s, includes Helm
Find out what Helm related talks will happen and put together a list in hackmd or somewhere
Chart caching, using it, and content addressability https://github.com/helm/helm/issues/4618 [Matt Farina]
Cache as-is doesn't work because could have same chart and version
Use digest instead as cache key for Helm 2
Homebrew Helm problem
Fixed and merged!
Justin: Docs site search experience improvement
Farina: Outstanding issue to move Helm docs out of tree to another repo
Assignments
Moderator: Adam
Notes: Fisher
Issue Sherpa: Justin
September 20, 2018
Announcements
Helm 2.11 RC4 released yesterday
1.11 Kubernetes support
https://github.com/helm/helm/tree/v2.11.0-rc.4
Standup
Justin


Matt Butcher
Putting together helm demos for ignite
Brian
Working on lua implementation
Adam
Values files stuff
CRD, replacing secrets
Tiller race conditions
Michelle
2.11 release
Matt Fisher
2.11 release
Adnan
Monocular 1.0
Cron jobs to sync chart repositories, more responsive
Matt F
Waiting for cluster, still being worked on
Policy, workflow and tooling stuff for the Hub
Docs for how to stand up a helm repo
Chart-testing images now up on quay
Josh
Helm push and authentication

Discussion
Helm -push multiple provider support
Something in index.yaml?
Governance
1 week left for nominations, 9 people nominated and accepted
7 initial selection, but can have up to 9
https://gist.github.com/mattfarina/51056f5655dea3f0d60e20b37ccc79c2
Using Quay for chart testing
https://quay.io/repository/helmpack/chart-testing
Provides usage logs
2.11 official release date
Bugs should be squashed
Tuesday Sep 25th (unless explosions, Wed)
Why Lua?
Need to explain to users the reason for this
Why not JS etc
Write a post/doc that dives into depth
Assignments
Moderator: Matt Fisher
Notes: Adnan
Issue Sherpa: Justin

September 13, 2018
Announcements
Helm 2.11.0-rc.2 was released yesterday
Helm Governance votes were complete last week
Standup
Michelle
Issue triage
Working on proposal for auto rollback feature
Fisher
Working on the Helm 2.11 release making sure that’s going through
Adam
Working on some of the Helm 3 design stuff with CRDs
Out of town next week so will miss next week’s standup
Matt Farina
Working on the charts repository stuff, DCO labeling
Throwing the charts-testing image on quay.io
Working on putting a blog post on working with the charts-testing project
Will be weighing in on the distributed search proposals
Brian
Working on the Lua VM implementation for Helm 3
Josh
Closed a few ChartMuseum chart PRs
Looking into authentication/authorization flow for Helm 3
Discussion
Chart packaging system improvements (@gintas) Please see https://hackmd.io/Q8dyCyztRmW4_Xz9UFwuMQ
Farina: what about the forking model? If you fork the chart and not change the ID (an implied change of this proposal), which cached version wins? You’ve got a non-reproducible cache pollution issue.
The question becomes “how do we do caching/forking with this model?”
We could run a hash on the chart and map it to that, perhaps
Josh: hashing may be a problem with relative file path links in requirements.yaml (charts can change on disk)
Farina: I’m not seeing what particular problems this proposal is trying to solve. Is there a human-readable write-up on the problem at-hand?
Action item: reach out and follow up with @gintas on the core problem without trying to come up with a solution
https://quay.io/helmpack/chart-testing/ is now public and we are archiving logs for long term via https://github.com/mattfarina/query-store-quay-logs (to be moved to the helm org after elections). But, only archiving aggregated stats (e.g., total pulls per day) rather than details because detailed as usernames mapped to IPs and other data. Thoughts?
Fisher: does the CNCF have any policies/thoughts on statistics for CNCF projects (e.g. any policies that came out of devstats.cncf.io)?
Farina: good question. I’ll follow up with them
Adam and Matt: probably good to just track general trendlines rather than deep personal data on users.
Josh: can we move the ChartMuseum image over to the quay repository?
Farina: yeah!
Travis Foster: hey, I’ve been looking into JSON parsers for Helm 3
https://github.com/helm/community/issues/49
Farina: thank you for taking a look into this! We’re looking for ways to looking into performance gains we can do with parsers so this is helpful.
Assignments
Moderator: Matt Farina
Notes: Michelle
Issue Sherpa: Matt Fisher
September 6, 2018
Announcements
Governance has enough votes to pass! Acceptance window closes today at 5 pm Pacific
Standup
Taylor
At Open Source Summit last week
2.11 PRs this week
Josh
Exploring Helm login and authorization flows. Been talking to Matt Farina and some of the people on the Microsoft Containers team
Adnan
Merged the Monocular 1.0 proposal
Matt Fisher
At Open Source Summit last week
Working on 2.11 PRs this week.
Matt Farina
Working on Governance this week
Charts testing image work. Not going to put it on Docker Hub and going to put it on Quay instead. Going to write a tool to pull the logs from Quay and upload to storage
Michelle
Working on proposal for auto-rollback when releases fail
Adam (vicariously through Michelle)
Working on Helm 3
Discussion
Helm 2.11 release
3 PRs left in the queue before we can cut a release, all ready to go: https://github.com/helm/helm/milestone/55
Should we pull https://github.com/helm/helm/pull/4468 off the milestone? (fisher)
We are ok with pushing this off to 2.12 so we can hammer out some design issues.
RE: PR #4590, was there a technical reason we use cmd.PersistentFlags() in certain commands vs cmd.Flags()? (fisher)
PersistentFlags used so all subcommands can inherit the flag. Fisher is going to go back and change this for some of the new code
“maven-like dependency system” for charts (josh via gintas)
org.apache:kafka:1.0.0 -> ~/.helm/repository/org/kafka/1.0.0/kafka-<classifier-id>-1.0.0.tar.gz
Possibly adding dependency caching
Josh is going to work with a few others to create an issue on this
(Henry Nash): As a newbie, have some questions about how the overall community works etc.
IBM is starting to use Helm a lot and wanted to get involved
Assignments
Moderator: Josh
Notes: Fisher
Issue Sherpa: Michelle
August 30, 2018
Announcements
Switched to the DCO from the CLA
Blog post here
Standup
Adnan
PR’ed Monocular 1.0 proposal
Matt Fisher
At Open Source Summit NA
Paired with Taylor around 2.11 release, triaging issues
Brian Hardock
Lua work for Helm 3.0
Michelle
Working solution for Issues #3805 and #1193
Vic
Worked with folks at Replicated Ship
Their business is to ship software to vendors
Did presentation at SIG Apps
Worked on how to do customization around helm charts for different environments. “Last-mile customizations”
Matt Farina
DCO change + prep work for DCO change
Follow up issues on charts automation
Made changes to governance doc
Matt Butcher
Worked on governance docs with Farina
Josh
Giving a short talk on helm 3 in SF
Adam
Working on refactoring for Helm 3. Got a bunch of stuff merged.
New API for dealing with charts and releases
Setting up for Lua stuff
Working on upgrade issue with Fisher and Michelle
Discussion
Estimating the Airspeed Velocity of an Unladen Swallow
Bumped to next week
Should we use something like https://github.com/ericchiang/k8s instead of client-go for helm v3? (Farina)
Client-go has some well known pain points
Butcher: is it still being maintained?
Not sure
Would it be practical to use an alternative client to client-go in helm 3?
Butcher tried it. His findings:
Pervasive change
Eric made some really insightful changes but that create a somewhat cumbersome experience
For example, pointer to all the fields
This is the approach that apache thrift took
To use this library, some things are more cumbersome for the user
Some parts of the library didn’t feel 100% done
Pros: dep ensure took like 5 seconds
Adam: what is driving this?
Farina: Making upgrades easier. This is the time to make breaking changes, so let’s entertain the idea.
Adam: will take a look at it this week
Docker Hub vs alternatives for public image distribution (Farina)
On the charts side, we are going to distribute the chart testing library. Been thinking about where is the right place to store it and point people to.
Went out on the email list
Butcher: DOCKERHUB
Quay will tell us if there are vulnerabilities
Michelle: security is nice. Cool.
Fisher: DockerHub seems to be ubiquitous
Deis Workflow had a tough time trying to figure out metrics like image download count. Didn’t know if it was actually accurate.
Metrics not a blocker if the discoverability is the primary driver for this.
What is the best place to store images to get metrics?
Metadata points in ACR but it’s not public so it’s not an option
A wild Taylor appears in the meeting
Is DCO required for a PR with a CLA that was submitted before the DCO took place?
Should not be required. PRs are covered before
Assignments
Moderator: Michelle
Notes: Taylor
Sherpa: Fisher

August 23, 2018
Announcements
Standup
Fisher
Monitoring 2.10 release and making sure everything is going as planned
Matt Butcher
Issue sherpa for this week, nothing significant to report
Working with Matt Farina on the governance proposal
Talking with Brian and Adam on some of the Helm 3/Lua stuff
Adam
Continuing work on refactoring Helm 3, looking into breaking things into smaller chunks to move things upstream
Merged Kubernetes 1.11 support into master and the dev-v3 branch
Michelle
Reviewing Helm PRs
Re-named the helm-secure plugin to helm-rbac
Taylor
Doing some PR review work
There is an open PR for recursive build option, code looks good but there may some eyes on it
Asked Matt Farina if he can help take a look at it
Adnan
Working on a proposal for changes to Monocular
Fisher
Released helm v2.10.0
Scouring the issue queue to make sure there’s any issues where we may need to cut a v2.10.1, so far doesn’t look like there’s any
Josh
Traveling etc last week
Working with helm repository authentication stuff
Planning to add “helm push” API spec to specs repo soon
Meeting w/ Zach (on Matt Farinas team) today about Chartmuseum bearer-auth PR and next steps. People have been asking for this more and more often. This will also tie into “helm login” proposal
Giving a talk on Helm 3 next week.. What do I need to know? Going off of Matt Butcher’s post (https://sweetcode.io/a-first-look-at-the-helm-3-plan/)
Chartmuseum website to launch possibly next week. Preview: https://codefresh-io.github.io/chartmuseum-www/
Matt Farina
Working on the governance model as well as the switch to the DCO
Justin
Won’t be available for 3 days
Someone at istio was being blocked with https://github.com/helm/helm/issues/3805
Adam: this would be a major breaking change to how Tiller works for Helm 2, but it could be a possibility for Helm 3
There are a ton of edge cases around the different merge strategies
Farina: what about server-side apply?
Adam: same problem, it would be a very large technical challenge to manage different upgrade strategies for Helm 2
Discussion
Question: how does helm plan to address auto-generated configuration on `helm install`? E.g. autogenerated passwords in the postgres chart
Farina: we should come up with a convention.
Butcher: I would guess doing a pre-install hook to install a Configmap to autogenerate a password and place it in there
DCO?
Butcher: DCOs require all commits need to be signed?
Adam and Farina: yes
https://github.com/mattfarina/brigade-test-project/pull/1
https://github.com/probot/dco
Farina: asking michelle if they can pair on this on Monday?
Michelle: yes
Helm v3 or hub: enforce or notify if immutability changed?
Farina: if the digest changes, should we notify if someone broke
This is referring to changing an existing chart in-place, not bumping the version
Michelle in chat: then yes, we should put a warning in
Adam: the one thing we may want to think about is when charts are in development.
Farina: we could do a digest diff on a `helm repo update`?
Farina: we should probably think this one through
“helm login” - How do we envision this to work? Anybody interested in a call to address this in detail?
Josh will put out an email to discuss this topic
Butcher: can we start on the next Helm release since k8s 1.11 support is done and we have a few minor features committed already?
Fisher: sure
Butcher: there’s a lot of people asking for it, so we should cut an RC
Taylor: we should go through the PR queue before we cut an RC as there’s been a few requests
Assignments
Moderator: adam
Notes: michelle
Sherpa: justin

August 16, 2018
Announcements
Slack is down, meeting link needs update.
Standup
Taylor
PR work, with follow-ups today and tomorrow
Matt Farina
Chart PR queue reduced. Now below 400
DCO chart automation is a blocker to widespread DCO adoption.
Distributed chart site is underway. Looking at hosting
Link for this meeting is wrong in Helm repo README. Will fix.
Butcher
Main focus right now is on Helm governance, CNCF deadline is September
Working with one of Microsoft’s engineers on a bug with `helm ls`
Next week: continue to focus on governance and focus on next release for Helm
Helm 3 related work (reporting for Adam): onto the Helm 3 loading and rendering logic
In-memory tree structure is structured rather than flat
Lua side (Brian): the core engine is about halfway there, working on Lua 5.3 support
Matt Fisher
Mostly issue queue and release management
Should the issue go in RC4 or are we good to go with 2.10.0?
Adnan
Changes to Monocular with the hub stuff
Nikhil
Out last week. Getting back into things.
Michelle
Mainly responding to GitHub issues
Will respond to comments on test proposal in the helm/community repo this week
Josh
PR out for ChartMuseum bearer auth
Exploring approaches for values management
Should we add ChartMuseum API for fetching/saving value sets?
Allow for diff key-value backends (Consul, Vault)
“helm config” plugin that works with API described above to fetch value sets and save the to .yaml file prior to helm install
Going to get involved in specs repo as it relates to helm push, helm login
ChartMuseum website coming soon - can we open source it under helm org?
Demo
Charthub (Paul Czarkowski)
Easy way to host chart-repos using GitHub releases.
[Adnan] may be interesting to use this as the base for the “github bot”
Discussion
Longer meeting? Should we start 30 min earlier?
Are we letting meetings run long because we don’t do well time-boxing items, or do we just have a lot of agenda items.
There seems to be more material to discuss. With Helm 3 and the hub, there is more “stuff” to discuss.
Should https://github.com/helm/helm/pull/4473 be thrown into the release branch and cut rc.4, or cut v2.10.0 without it?
[Taylor] would prefer to see a 2.10.1 with this fix in it
[Fisher] would like to cut a 2.10.0 today
Assignments
Moderator: Farina
Notes: Fisher
Sherpa: Butcher

August 9, 2018
Announcements
Helm v2.10.0-rc.3 was released yesterday
Standup
Butcher
Staying on top of issues
Research for Helm v3 command line UX, PR in helm/community repo
Suggested a few edits
See doc for analysis
Turning attention back to governance. There is a draft out there.
Looking at Lua things
Brian Hardock
Building a Lua implementation for Helm v3
Focused on the virtual machine / interpreter for Lua and writing the backend
Had conversations on design with Butcher
Fisher
Majority of work has been related to the release candidate
Helm issue #1968 - meta issue on the build system and the infrastructure around it
Moderate discussion on building helm for different architectures
Continuing work on release candidate and watching for regression and going forward with that process
Taylor
Clearing out the PR queue as much as possible
Continuing clearing out the PRs
We are not currently accepting PRs against the helm3-dev branch
There is a massive internal core refactor which is why we’re holding off on merging any PRs (Farina)
Farina
Kubernetes came out with recommended labels similar to our own and did a PR to update our docs with that
Following up on governance related items
Following up on charts

Adnan
Getting next release of KubeApps.com out
Adam
Heads down on refactor work for v3. Close to having the public API part down for helm v3.
Watching for feedback on the PR for supporting kubernetes 1.11 work
(Taylor) Doing more testing against AKS right now
Vic
Plugging back in
Will work on Kubernetes 1.11 verification
Working on figuring out how to ease use cases observed and pain points from helm 3
Michelle
Reviewed a bunch of PRs, especially patches for release candidate
PR’ed helm test plan/proposal for helm v3
Reviewing long term governance plan
Josh
Added Helm 3 proposal for helm push
Trying to tackle repo authentication design w/ Farina and friends
Helm-push plugin updates (Harbor project etc)
CNCF Webinar w Stef! https://www.youtube.com/watch?v=u3VqswB-TJo&feature=youtu.be
Discussion
How are the changes/testing coming along for kubernetes 1.11 support? Where do we need more help? (michelle)
Coming along well. Adam already updated.
Please help if you can.
Specs repo, helm push/login
Clear specification for authentication and uploads by default
ChartMuseum will be responsible for implementing this spec, but others can BYO-API
How do we does the auths
Will pull in JFrog and some other people for their experience
(Matt Farina): Would like to get the specs repo moving. Being voted on now. Especially important for distributed search.
Hub update? Is there any movement on putting forth a monocular instance to start the hub
(Farina): yes there is movement
Need to find location where it will be hosted
CNCF has a bare metal as a service option
Few other ways we can go between sponsored hosting from a company or through the CNCF
Farina is chasing down funding
Guidelines for inclusion
What guidelines should be there to be included
Reinhard (chart maintainer) is working on this
CNCF will look at this as well
Adnan will also help because Bitnami has experience in managing chart repositories
There was an email sent out to the maintainers to figure out a name
Adnan has been looking into resource requirements for hosting monocular
Very little resources required
Guides about creating your own repositories
Started with how to host your repo on GitHub
Key focus on documentation
This is distributed search not proxying for repos
First version will be a monocular instance and we will improve from there
Assignments
Issue Sherpa: Matt Fisher
Notes: Josh Dolitsky
Moderator: Taylor

August 2, 2018
Announcements
Helm RC2 was released earlier this week. RC
Standup
Nikhil
Getting back from an internal all-hands
Will be testing out the release candidates
Adam
Opened a PR for k8s 1.11 support, could use heavy testing. Testing different k8s versions with it. Wait testing. Anyone with cycles who could help will be appreciated
Other Helm3 stuff
Brian
LUA stuff for Helm3
Fisher
Making sure helm2 RCs are going through, move along the 2.10 cycle
Paired with Adam for the k8s 1.11 support.
Keep moving through the RC2 => RC3, but has already been approved, just needs merge.
Justin
Some issue triage this week
Butcher
On vacation last week
Getting close on the Helm Governance Proposal -- should have something on the next couple of weeks. (deadline in Sept)
Working with Ronan to get various helm accounts shifted over to CNCF
Item on to-do list to figure out where to move our docker image to. Also an open issue on where to host our helm binaries. The crucial thing here is metrics -- we’d like to host it where we’d be able to get rich metrics around downloads. Need to co-operate with k8s here
Adnan
Working with Scott Rigby to figure out how to make standing up distributed chart repos as easy as possible.
More on this soon
Discussion
Update strategy in helm v3. See issue #3805 (michelle)
Adam’s been working a bit on a prototype to evaluate this. Pretty close to spiking it out, when he should have a better idea about this.
`--strict` flag on `helm repo update` to return non-zero exit code on failure PR #4348
<Hard to figure context on this, so let’s discuss in helm-dev>
Release Helm 2.10.0 today after merging PR #4364? (fisher)
Already discussed cutting an RC3 before doing a 2.10.0
Release without Adam’s patch for k8s 1.11. Doing so would break semver, so we shouldn’t do this.
Proposal to port GoFish as a Helm 3 plugin manager (fisher)
https://github.com/helm/community/blob/master/helm-v3/005-plugins.md
Is the community repo the right place for this?
Yes!
When we have something in an issue queue that’s possibly a bug, currently we mark it as question/support. Can we have a new tag “Unconfirmed” where we can mark it as an unconfirmed bug, and then remove the tag?
Yes
https://github.com/kubernetes/kubernetes/labels/priority%2Fawaiting-more-evidence
Helm v3 update:
Ongoing work on the LUA engine
Proposals in helm/community are flowing, will spend some time next couple of weeks going through them.
No definitive ETA yet, but it would be nice to be able to show some of the newer features during KubeCon
(Adnan) Vic is holding a discussion around last-mile customisations to charts in next Monday's SIG Apps call. Also Stefan Prodan from WeaveWorks will be demoing Helm git-ops.


July 26, 2018
Announcements
Finally finished move of all the helm repos to the Helm org
Distributed search proposal passed with a majority of votes
Voted Rimus (sp?) as an emeritus member
Standup
Butcher
Issue sherpa trying to stay on top of issue queue
Merged PRs -- were a couple of features
Helm 3 design LUA stuff with Brian
Adam
Been working on getting the 1.11 kubernetes support ready
Josh
Been trying to keep up with email list
Been a lot of activity around repo discussions:
Taylor
Vacation
Working on PR stuff tomorrow
Justin
Was at vacation
Getting back into the flow of things
Farina:
Need to figure out logistics of moving from CLA to DCO
When we left k8s we lost auto-sizing, have been figuring out how to bring the sizing back. Chris is going to try and chase down how to use ProBot so that we can get some of this back.
Working on update notifications -- how to tell folks that they have a helm update for them.
Discussion
Can folks (non core-maintainers) add/remove labels on the helm/helm repo right now? (michelle)
Non-core shouldn’t have access to do this.
There used to be a bot that we could use. The automation in place used to be able to allow folks to do this, but it was k8s specific.
Helm proposals
Installing custom plugin sets:
(Taylor) Two thumbs up
How do we do version handling? How do we discover the version.
(Butcher) It might be good to not even handle versions for a first pass
Helm v2 repositories
https://github.com/helm/community/pull/27
Two file approach -- idea is to do this to improve performance so that we’re able to do a lazy fetch
If you use a relative URL, don’t you lose context of the base URL?
If you have it in more than one URL the relative URL works across http / https.
Are we going to host keywords / maintainers from the current charts repo into the top-level file? How do we consider what is latest? (eg. 5.1.3 vs 5.1.4-alpha1).
We might want to change it from “latest” to “stable” to address this
Improved signing
https://github.com/helm/community/pull/20
Generate GPG keys for you so that you can sign charts (default is on). Move closer to making verification a part of the helm repo experience.
Main thing here is whether storing the public key is okay to store in the charts or not.
We don’t need to take an external dependency on GPG -- go already has libs that we use (already in the code).
Update notifications
<Punted to the next meeting in the interest of time>
Stale issue automation
<Punted to the next meeting in the interest of time>
Assignments
Issue Sherpa: Justin
Notes: Nikhil
Moderator: Butcher

July 19, 2018
Announcements
Need helm doing some research into possible bugs from the issue queue. If you have time today, please reach out to Michelle
Accelerated move to Helm org (github.com/helm/helm)
Still having some CLA/CicleCI issues
Seems like it might be an issue with CLA-Bot
Standup:
<Skipping standup so that we can devote some talking about the Governance changes>
Discussion
Governance
Matt Farina's proposal for a new governance structure
https://hackmd.io/s/B1BksVYW7#
Bootstrapping process for project maintainer elections
Still need to figure out how to do this fairly and equally
Project maintainers need to be codebase maintainers
A codebase maintainer may not want to/be suitable to be a project maintainer
Codebase maintainers have skin in the game and care about the project direction
Codebase maintainers don't need to be software engineers, Helm has a history of maintainers that didn't contribute software but have contributed in other ways
Experience matters here, want ideas and psychology to bubble up to project maintainership
Don't want to end up with top-down maintainership
Sub-projects have a lot of freedom in adding maintainers - simple majority vote
Action: Make it more explicit that codebase maintainers don't need to be software engineers
Okay to have an even number of project maintainers for a short amount of time, if someone steps down before replacement found
Maintainers only removed if they become unresponsive
Bootstrapping project maintainers
Different weights for different projects?
Ghost period for code maintainers should be shorter?
Set an upper limit, but let codebases choose lower limits if desired
CNCF CoC - email contact in the CoC doc
Better nomenclature for project & codebase maintainers - repo & org maintainers?
Chartmuseum
UI
Orgs and Authz
Assignments
Issue Sherpa: Butcher
Notes: Nikhil
Moderator: Fisher

July 12, 2018
Announcements
<None>
Standup
Taylor:
Getting back into doing Helm stuff.
Digging into PRs and notifications
Then looking into Helm 3 stuff
Adnan:
Discussions around charts - especially distributed repository proposal.
There were some ideas in the charts call this week which is recorded.
More conversations with Scott Rigby
Work with Kube CRD and Tiller Proxy
Nikhil:
Not much going on with Helm this last week
Just touching base with issue comments and PRs
Michelle:
Knowledge transfer with Adam on the changes for k8s 1.11
Working on the proposal for what Helm tests should look like for Helm v3. Looking at issue queue for what people were requesting. Looking to get the proposal
Brian:
Fix the 1.10 RC for helm list,
Looking into the TLS security support question. Turned out to be a documentation issue
Farina:
Notifications: Working on how we can notify people when new versions of Helm are out. Using mini-kube as a model for this.
Verification work with the proxy (using kubeapps model). This is interesting especially with the multi-tiller approach that most people use.
Lots of emails around governance
Discussion
Moving docs [Matt Farina]
Docs repo?
Currently docs live in the helm repo itself.
Would be nice to add translation, etc.
Keeping this in mind does it make sense to move docs to their own repo?
Markdown vs Asciidoc:
Conversation with a few other folks about this:
Markdown is more popular, asciidoc is more technical?
There are tools that can switch between the two.
Example of kube-builder http://book.kubebuilder.io
Nice to be able to do side-by-side code and docs.
Built in i18n with gitbook
[Taylor]: In support of moving the docs to own repo
[Michelle] Helm v2 vs v3 docs -- will they live in the same repo in different directories?
Docker compose does a good job of splitting out the docs for different versions.
Gitbook vs other tool? https://github.com/kubernetes/helm/issues/4100
Multi-lingual
K8s recommended labels. Do we switch? When? https://github.com/kubernetes/website/pull/8703 [Matt Farina]
Helm has standard labels, almost one:one with the ones listed here, but having different names.
What is the best way to proceed here?
Adding annotations makes sense, but labels changes people’s deployments so the change is more intensive.
Governance esp w.r.t related projects:
Separation of concerns project maintainers vs. codebase maintainers
CNCF projects are expected to follow standard set of CI / Automation practices. (https://bestpractices.coreinfrastructure.org/en) We will probably need to get to a certain state on this in the future.
Assignments
Issue sherpa: Michelle
Notes: Adnan
Meeting Moderator: Nikhil
July 5, 2018
Announcements
Devoting time to discussion, skipping standups
Discussion
Discussion on k8s 1.11 support for Helm
Update on current state of k8s 1.11 backwards incompatibilities
Generated Go code in k8s/k8s has been removed, making it unusable as a package: https://github.com/kubernetes/kubernetes/pull/62432
Removed the client-side reaper in favour of the server-side reaper. Because Helm prior releases, we need that for 1.10 support
The entire factory/builder packages in kubernetes/kubernetes were rewritten and are totally different between 1.10 and 1.11
The printer was rewritten and requires significant refactoring
Proposal (adam): due to backwards incompatibility and significant engineering effort, I propose Helm 2 only supports Kubernetes 1.10, punting k8s 1.11 to Helm 3
Another option: remove k8s/k8s as a dependency, and vendor the k8s 1.10 code directly into Helm 2 as a crutch
Apparently service catalog has already gone with this approach
Butcher: (giving some context on kube-lego and cert-manager) I don’t think we should go with solution 1. It puts our community in a bad position: Helm 3 isn’t ready, and Helm 2 would be in limbo
Adam: ok, I can go take a look at forking kubernetes/kubernetes so we have the packages for handling backwards compatibility
Farina: I’m going to take these issues up to SIG Architecture and SIG Release
Michelle (in chat): you think we should bring this up in SIG apps and see what other tools might be struggling with changes from this kubernetes release?
Farina: on the agenda for next SIG Apps
Fisher: Do we want to have a strategy for communication with the community for this? Farina: Better to over-communicate than under.
Adam: Go/no-go by end of the week. Farina: based on this we can craft some messaging to the community next week.
Assignments
Issue sherpa: Justin
Notes: Nikhil
Meeting Moderator: Michelle
June 28, 2018
Announcements
v2.10.0-rc.1 released
Standup
Justin:
Issue Sherpaing
Working on CRD issue
Adam:
Released v2.10.0-rc1
Still been working on some of the Helm 3 stuff
Pairing with Brian on some things to help through some of the design choices
Working on k8s 1.11 support into Helm, but it’s so vastly different than 1.10 so it was unable to make it into the release
Brian:
Paired with Adam, and I loved it!
Josh:
Nothing to report
Adnan:
Nothing to report
Matt Fisher:
On vacation last week
Testing RC.1
Discussion
Kubernetes v1.11 support
Assignments:
Issue sherpa: Justin
Notes: Matt Fisher
Meeting Moderator: Adam

June 21, 2018
Announcements
<None this week>
Standup
Nikhil:
Issue sherpa last week - support related things
Talk about cutting a new release?
Matt Farina:
Migration to Helm org
Michelle opened up an issue moving from CLA to DCO that needs follow up.
Some work on index files and performance - splitting up files and moving to JSON could yield a lot of benefits so exploring those options. Upcoming v2 proposal spec around this.
Adam:
Out sick last week.
Being pulled into other things since Butcher’s out.
Helping with Helm 3 refactoring
Fisher:
Issue queue + triage
Had a good session with Nik helping with Issue Sherpa handoff, and process regarding labels while sherpa’ing
Josh Dolitsky
Keeping up with all the stuff Matt Farina’s been posting with distributed search
Wants a heads up on any CNCF stuff going on
Farina mentioned about copyright headers
Question: is there any hold-up on moving to the github org
Farina: one hold-up is the Linux CLA bot hasn’t; once it’s fixed we can start moving over
Adnan
Been looking over the distributed search proposal and the application CRD
Justin Scott
Been heads down on internal projects, LGTM’d a few PRs
Discussion
Next helm release?
Talk about cutting a helm 2.10 with the changes present in master
K8s 1.11 client APIs broke a lot of things from initial testing
Come up with an announcement on the helm-users mailing list on if we’re adding 1.11 support
Helm 3 tickets
There’s been a few tickets raised on Helm 3 topics; are we ready for these?
Answer: no, redirect users to #helm-dev or the helm/community proposals at this point
Things are expected to break (and break often); we’re not ready feedback until the first alpha release
Zoom issues - people on hold?
The meeting moderator now needed to allow folks explicitly into the meeting, and folks are on hold until the moderator allows them in
Farina: Default k8s settings for zoom are changing to make things much more restrictive, and moderators need to explicitly let people in. This is around the push to stop bad actors, but meeting chairs need to keep an eye out for folks trying to join.
It might be effective to have co-hosts so that we can have multiple people keeping an eye out to let people on hold to join in.
Assignments:
Issue sherpa: Justin
Notes: Adam
Meeting Moderator: Fisher

June 14, 2018
Announcements
<None this week>
Standup
Fisher:
Was at offsite last week
Working mostly on Issue queue / PRs
Fix for: https://github.com/kubernetes/helm/pull/4146
Adnan:
Need to check out Helm v3, and got feedback on Helm CNCF move
At DockerCon this week
Nikhil
Taking a look at the helm 3 branch, looking at particular the lua stuff
Adam
Out for last couple of weeks
Looking into some pieces of helm that weren’t touched before (requirements)
Need to look at treatment of Helm subcharts internally (needed for LUA)
Butcher
Pairing with Adam on the tech side for Helm v3 -- taking a deep pass through the lower code base before building the upper layers
Will be out next couple of weeks (including at ContainerCon China)
CNCF Helm transition is going well, but slowly -- turning over various properties (web, DNS, Twitter handle, etc). Put in place a provisional Helm governance model. Provisional governance structure is for 90 days until we can come up with something more formal and structured. Things that have moved over officially:
Helm
Charts
Chart testing tools
ChartMusuem
Everyday conversation will stay in Kubernetes Slack.
Will be running an article in https://sweetcode.io/ on Helm 3
Taylor
Not much on Helm, mostly working on AKS GA
Hope to jump back in after next couple of weeks
Discussion
How can we better manage issue triaging with the issue queue getting so large?
Something that came up last week. Currently ~400 issues, ~90 open PRs. What’s our level of priority for issues / PRs coming in.
4 categories: Bugs, Support Tickets, Features, Whatever else?
59 open bugs (most are fairly low pri), but there are a couple of high pri ones
Thought is: make handling bugs as P1 for issue sherpas since it’s hard to focus on everything.
(Adam) Another idea: talk about high-pri bugs during meeting.
Charts distributed repo proposal - please leave feedback https://github.com/kubernetes-helm/community/pull/24

Assignments
Issue sherpa: Nikhil
Notes: Fisher
Meeting moderator: Adnan
June 7, 2018
Announcements
<None this week>
Standup
Taylor:
Not much this week
Discovered issue with --wait
Nikhil
Issue sherpa, spent some time on some issues and looked into a few PRs
Adnan
Plan to take a look at the Helm v3 branch, and the canary
Not much else apart from that
Discussion
Stef: Issue with “Connection Reset” caused by difference of versions between Helm and Tiller
Assignments
Notes: Taylor
Meeting Moderator: Nikhil
Issue Sherpa: Fisher

May 31, 2018
Announcements
<None this week>
Standup
Matt Fisher
Issue triage + PR
Some windows related stuff
Brian
Nothing to report
Taylor
Trying to make time to fix the helm wait stuff
Justin
Nothing to report
Nikhil
Started pulling down on the dev-v3 branch and making his way through the branch to see what’s changed
Adnan
Interested to see what’s going on in the dev-v3 branch, otherwise working on other work
Matt Farina
Worked on the helm proposal to the CNCF, fair amount of discussion going on in that PR. Discussion is still ongoing
The community charts sync broke due to a breaking change in the system
Fixed by using CircleCI rather than Jenkins thanks with help by Adnan
Discussion:
Folks are in summer mode! Not much tabled to discuss this week.
Assignments
Moderator: Taylor Thomas
Issue Sherpa: Nikhil
Note taking: Matt Fisher
May 24, 2018
Announcements
Helm and CNCF [Matt Farina]
After last week’s incident, we are going to be stricter with how we can handle this in the k8s community. Sorry to all about last week’s incident, but we should now be in a better place to handle something like this if it happens on a CNCF community call again.
Standup
Taylor:
Not much, mostly working on the wait-flag stuff
Might be busy over the next couple of weeks
Adam
Working on Helm 3 stuff
Trying to split up the tight coupling of templating and resources
Next week on vacation
Nikhil
Mostly out last week, getting back into the swing of things this week
Butcher
PR on the LUA stuff for Helm 3 (Core event handler)
Talked with Adam about things like helm install which has grown organically -- so we need to do a little ‘topiary’.
Jdolitsky
Managed Helm Repos
We also now have auth for ChartMuseum using a bearer-token in the header
Worked on helm-push/helm-login/performance of index.yaml
Farina
Happy to take on the proposal of the index.yaml
Starting to poke at i18n, filed a few issues against doc, i18n, etc. Looking into finding a framework to do the translation.
Fisher
Mostly on vacation last week - got back this week
Hoping to take a look at Helm3 discussions and get involved with that this week
Discussion
(Taylor) Wait-flag returning true if the app can’t start. Not sure if it’s worth it to keep the --wait flag around. Want to get input from community, and maintainers on whether it is a good idea to remove this. (Farina) On the community charts side, they had to write their own solution, but something like the wait flag is a better solution. Do understand the brittle nature of it, but it is useful. (Adam) There is a effort upstream that we might be able to piggy-back on (https://github.com/kubernetes/kubernetes/pull/64034)  -- so we might be able to leverage some of that. Maybe we should review what’s going on there, and get caught up on discussions. (Fisher) What’s the status of wait flag on v1? (Farina) There’s a storage bug that’s preventing old API versions from being removed. Hopefully that will be solved so that we don’t need to support older versions of the API / deployments? Probably a good idea to see in what direction the upstream effort is going before making an informed call here.
Josh Demo https://github.com/chartmuseum/helm-push
Assignments
Issue Maharaj
Butcher
Notes
Fisher
Meeting Moderator
Nikhil
May 17, 2018
Announcements
Helm and CNCF [Matt Farina]
Standup
Taylor
Debug --wait flag
Opened PR for release process
Nikhil
Issue sherpa duty
Fisher
Vacation
Defining first issues with Farina
Adam
Helm3 refactoring
Review PR’s
Michelle
Review PR’s
Butcher:
Vacation
PR reviews with Adam
PR for CRD support in helm2
Farina
Looking into making chart repo more distributed
Language support for helm
Demo
Josh with Chartmuseum
Discussion
Kubernetes Apps Survey Results for Helm [Matt Farina]
Issues with the wait flag [Taylor]
Assignments
Issue/PR sherpa:
Michelle
Notes
Nikhil
Meeting Moderator
Fisher

May 10, 2018
Announcements
3.0 dev branch is sorta open but in code cleanup phase
Codefresh just announced managed Helm repos -- awesome!
Standup
Nikhil
Has not done much in the last week
Volunteered to do issue/pr sherpa
Reviewed oauth work with Josh
Adam
Holding fort when folks were at KubeCon
Focusing on Helm3 stuff with refactoring
Farina
Digging into the charts craziness
Almost 300 open PRs
Added a discussion topic to figure out how we can manage some of this craziness
Hope to chat with Josh and present a PR to help with this
Butcher
Picked up some of the Issue Sherpa this week
At KubeCon -- spoke to a few Chart maintainers to talk about some of the alternatives to what we have  currently.
Propose people interested in this problem, meet up during Chart Office Hours (next week on Tuesday?)
Open PRs for CRDs that needs reviews -- might need to pair with someone
Has a POC for LUA scripting for Helm v3
Taylor
Proposal idea for how we want to do releases
Looking into PR templates / Issue Templates
Discussion
Proposal idea for how we want to do releases [JTT]
The dev call after the release is the next (bootstrap) call for the next release.
Each dev call have a 5/10 minute discussion for features where we discuss whether we pull merged PRs into the next release of a later release.
Is this adding unnecessary process which will be overhead for us?
Taylor will write up the process and make a PR to Helm to kick off more discussion for next week.
Distributed charts vs central hosting vs a combo [Matt Farina]
Charts Office Hours are at 9am PST 10 am Mountain on Tuesdays
Whole lot of work to curate charts and get people to curate charts and maintain. People tend to get burnt out. So the central repo model doesn’t really scale that well.
PHP packages have distributed hosting -- can we have some sort of distributed mechanism to do a similar thing for charts that is distributed, but has a central discovery -- so central metadata, with the actual charts living off in a different location.
More discussion next week
That _is_ a different project, so technically they would make a decision on this, but we should weigh-in during Charts office hours
Assignments
Issue/PR sherpa:
Nikhil
Notes:
Adam
Meeting Moderator:
Taylor

May 4, 2018
Canceled due to KubeCon

April 26, 2018 [recording]
Announcements
Helm 2.9.0-rc5 was released last afternoon
Helm dev call canceled for next week
Standup
Adam
Been doing work on Helm 3 refactoring, pulling out gRPC and old commands we had
Paired with Fisher on some bugs for the Helm 2 release and issues upon request
Taylor
Haven’t done much, switched over to Microsoft so have been onboarding onto his team over there
Matt Farina
Been looking onto the community charts scalability problem
Looking into chart repository scalability issues, like switching to JSON for index files
Might have stronger opinions after KubeCon
Matt Butcher
Been issue sherpa this week, triaging issues for Helm 2
Pairing with Adam on Helm 3 stuff
Automated helm.sh and docs.helm.sh this week
Have a couple items to work on for the Helm 3 spec doc while in Copenhagen
Justin Scott
Issue triaging
Identified an issue around import-values, will look further into the ticket
Michelle
Been travelling this week
Matt Fisher
Been cutting release candidates as needed this week
Want to cut a 2.9.0 by the end of this week, but we should probably discuss this
Demo
Josh with push to chartmuseum
Discussion
Helm material from Kubernetes Apps Survey
More Helm Maintainers for the issue queue?
Should we cut Helm 2.9.0 prior to KubeCon?
Testing coverage (Adam)
Assignments
Issue Sherpa
Adam Reese, party of 1
PR Sherpas
Adam Reese, party of 1
Notes
Matt Farina (for the week after KubeCon)
Moderator
Matt Fisher (for the week after KubeCon)
April 19, 2018 [recording]
Announcements
Application survey date extended until Monday - https://docs.google.com/forms/d/e/1FAIpQLSeetP_EnzMjAayR73sbpHDtfxnAzT6Y0wzCCvrP4C1uLa1u0Q/viewform

Standup
Michelle
Issue sherpaing and fixing bugs
Matt Fisher
Worked on fixing bugs
There’s a PR to add flags to pkg that was merged and then reverted because it broke compatibility
Working on getting RC4 out and fixing bugs to do that
Nikhil
Didn’t do much Helm related this week
Farina
Not much directly related with Helm
If anyone wants to work on Dashboard (SIG-UI), they’re looking for folks (Currently Angular 1 with PR to update it). Low barrier to entry.
Butcher
Reviewing PRs for Helm 2
One PR open on spec for Helm 3 around on security
Going to proposes changes to where releases are stored
Adam
A few patches in for the RCs
Working on Helm 3 looking to factor out gRPC and remove tiller binary
Looking for feature parity with a single binary
Brian
Paired with Adam on helm 3 refactor
Josh Dolitsky
Built a plugin for chartmuseum push
Discussion
Issue Sherpa’ing with a second person was really good -- maybe we should institute a buddy system
When do we open up ‘the branch’ (referring to v3 branch)?
Few things to consider - How to handle feature requests coming in.
Bugfixes, do we need to cherry-pick bugfixes into both? How do we deal with this since it looks pretty different (eg. no longer protobufs, no longer generated code, etc)
How / when do we define a freeze for Helm v2 features?
Lots of folks still using Helm 2
(Fisher) Maybe have a window (of a couple weeks or so) where we allow features for Helm 2, after which we only allow features in Helm 3. (Farina) Maybe have a cutoff date which we tell contributors about?
(Michelle) Some folks might want to continue with Helm v2 and not Helm v3.
Adam is pointing out concern for the end user vs concern for contributors.
Michelle is in favor of a helm 2 feature freeze date. Will use labels to track incoming features to make sure they get mapped to helm 3
“1. Window for getting in feature requests before the v3 requirement
2. Labeling features in Helm 2 that we need to forward-port, so we can track those”
Issue sherpas can respond to Helm 3 branch PR requests appropriately.
Let’s schedule Josh’s demo for next week, for lack of time.
Assignments
Issue Sherpa:
Matt Butcher
PR Sherpas:


Notes:
Fisher
Moderator:
Michelle Noorali

April 12, 2018 [recording]
Announcements
Application Survey https://docs.google.com/forms/d/e/1FAIpQLSeetP_EnzMjAayR73sbpHDtfxnAzT6Y0wzCCvrP4C1uLa1u0Q/viewform
All results will be publicly available
Standup
Taylor
Reviewing PRs
Still need some help reviewing PRs
Justin
Put some labels on issues
Matt Fisher
Working on release
Influx of stuff coming in which is great because we’re getting feedback on release candidates
Issue sherpa - aggressively closing issues
Common issues with bare metal clusters. Maybe there is something address in upstream Kubernetes documentation
Still triaging issues for bugs related to RC
Matt Butcher
Operating Kubernetes clusters with aid from hat
Some light reviewing on PRs
Added to Helm 3 proposal some things about strengthening signature verifications. Make sane security defaults for signatures. Will make repositories do some leg work to set up signing/security.
Adam Reese
Sick.
Working on PoC work around Helm 3 proposal.
Looking at how to refactor out some gRPC stuff to make easier transition to client only architecture
Michelle
Following up on issues
Sent request in for
Brian Hardock
Not too much
Matt Farina
Comparing memory allocation of YAML/JSON parsers
TLDR; we’re using the worst.
Will post result
Might be worthwhile looking at a better YAML parser
TLDR from chat; Matt Farina is a nerd and everyone loves it
Josh Dolitsky
JWT validation in chart museum working
Helm push plugin working on oauth flow
Discussion
RC 4 from master? (Fisher)
Lots of activity has happened since the release of the first  2.9.0 RC
There’s a lot of features released in there
Should we cut from master with changes included and features or should we continue to cherry-pick?
Adam: let’s follow semver
Matt Farina: we have lots of silent users so let’s now throw them off
Michelle: Agree with Adam and cool with doing a quick 2.10 release with new features
Taylor: If people want the latest, they can use canary
In agreement on how to go forward. Cherry-pick and no new features.
Action Item: Let’s make that clearer in documentation either in FAQ or CONTRIBUTING.md
Assignments
Issue Sherpa:
Michelle Noorali
PR Sherpas:
Matt Fisher
part-time sherpas
Matt Butcher
Taylor Thomas
Adam Reese
Notes:
Matt Farina
Moderator:
Adam Reese

April 5, 2018
Announcements
The AppDef WG has put out a survey at https://docs.google.com/forms/d/e/1FAIpQLSeetP_EnzMjAayR73sbpHDtfxnAzT6Y0wzCCvrP4C1uLa1u0Q/viewform
Helm 2.9.0 rc1 and rc2 were released this week (thanks Farina and Fisher!). The notification
Standup
Taylor
PRs this week -- trying to get stuff merged
We now have a release branch, so we don’t need to wait for the RC cycle to be done
Michelle
Was issue sherpa this week, mostly just triaging issues and fixing bugs
Have several items for discussion at the end of this call
Nikhil
Couple good convos with Josh on ChartMuseum auth stuff
Adnan
Pass
Fisher
Triaging issues, helping out with release
Couple of bugs with the release checklist. Opened an issue yesterday to look into.
If there is an issue that needs to be picked into a release, please use the ‘needs-pick’ tag where x is the release.
Josh
New chartmuseum release this week. 0.5.1 which lets you serve multiple chartmuseum repos
Putting together notes on helm push as next steps
Discussion
How do we respond to semver issues like this (Michelle)
If you use the kubeversion field and you have a constraint like >=1.8.0 this becomes incompatible with GKE’s versioning.
How do we respond / have we responded to these kinds of issues in the past.
One of the solutions was to add something like -r0 and making it pre-release (Farina might know more about this option -- this might be something that CoreOS/Rancher do)
There are over 300 open issues and over 80 open PRs. Should we do a week or two of having two sherpas to help clean out the backlog after 2.9.0 goes out? (Michelle)
(Taylor) Issues might not be much of a problem -- but we should definitely look into getting people to jump on and close PRs. The issue queue tends to self manage especially with fetja-bot and us noticing when things go stale.
(Fisher) It would be nice to make the community a bit more self-sustainable. Fully agree with prioritizing PRs over issue queue.
(Michelle) Do we want to have a dedicated person looking at PRs?
Taylor volunteered, but was decided not to make it a formal role just yet.
Vote on combined maintainers mailing list here (Michelle)
Two options right now (suggestions welcome!):
charts-and-helm-maintainers@lists.cncf.io
helm-cartographers@lists.cncf.io
(Adnan) Discussion on charts collecting analytics - https://github.com/kubernetes/charts/issues/4697
Heads up that this discussion is happening - folks interested please join in at the next charts call where this will probably come up.
(Fisher) One considerations is that Helm install doesn’t have a really good way of notifying the users that some analytics are being enabled -- this is especially true especially in the case of dependent charts. (Adnan) The dependency chart might be a problem since the parent chart could set the on/off flag and users might not necessarily know about it.
(Fisher) It would be nice to make it blatantly obvious that we were turning on analytics (eg. homebrew, etc.).
(Kensey) CoreOS has one flag for all projects to turn off analytics -- would be good to standardize
Assignments
Issue Sherpa:
Matt Fisher
PR Sherpas:
Taylor
Nikhil
Notes:
Michelle
Moderator:
Taylor

March 29, 2018
Announcements
Bacongobbler: Adam, Michelle, Fisher tested Kubernetes 1.10, so it is merged. When should we cut the Helm 2.9 release?
Standup
Adam:
I am going to miss standup, my 2 updates: merged kube 1.10 support and worked on a POC of the helm3 ext model.
As always, dont trust anything @mattfarina says about me
Michelle:
Testing of 1.10
Investigating another issue #3655
Nik:
Looking at Docker Registry token format for registry planning for Chart Museum
Investigating how PIP/pypy do things (and launchpad)
Justin:
Helm issue queue triaging
Butcher
Helm v3 work + prototyping with Adam
Embedding LUA, eliminating grpc -- get clear on some of the areas for the Helm v3 proposal
Taylor
Working on a few PRs this week
Matt Farina
Helm Survey for end users
Adnan
Charts work on owners/maintainers
Check if any chart maintainers want to go emeritus
Brian Hardock
Fix on Helm List due to gRPC threshold, working on tests now
Matt Fisher
Testing 1.10 and issue triaging
 Also a little Lua experimentation just to get comfortable with embedding Lua in Go
Discussion
What do we want to name the mailing list for both charts maintainers and helm maintainers? Is it public or private? [Michelle]
Think about this one, we can decide further next week.
CNCF landscape [Matt Farina]
Uploaders for Plugins [Josh D]
How to implement ‘helm push’
Farina: I was thinking of doing it over http/https downloaders, so that plugins can implement custom push uploaders.
Josh: See #18 on the community repo, which could also provide indexers to tackle search scale issues, maybe others.
Farina: Some survey questions are relevant to this. I’d like to keep push and index separate so that we might be able to do push in Helm 2. Index is more complex/debatable.
Josh: Maybe make chart museum method the default, and allow plugins to override
Fisher: jFrog is also interested, maybe in basic auth because that’s the only available kind
Farina: Maybe we should ship with oauth support and token auth too
Josh: What if we don’t have a default implementation?
Call for next Monday after sig apps: https://calendar.google.com/event?action=TEMPLATE&tmeid=NDA3a2NsbzlxaWdpcnNuMm41NzNsdnFqaDggY29kZWZyZXNoLmlvX3BmbmlqajVrbjFmYzhqMnJtMXYwbWJoaTIwQGc&tmsrc=codefresh.io_pfnijj5kn1fc8j2rm1v0mbhi20%40group.calendar.google.com
TODO: Schedule call -- Josh Dolitsky
Branch naming for Helm 3 [Butcher]
Tentative plan is to have a place for Helm v3 code to land prior to KubeCon EU (beginning of May).
We had some ideas on how we want to branch for v3. Farina and Butcher looked at data and found a substantial amount of folks cloning helm master.
Normal weekday 150 - 180 unique cloners.
Points to not breaking master and leaving as is. => v3 code should live in a different branch.
‘v3’ is as good a name as any for the branch name.
Farina: We should think about how the branch name plays with dependency management tools like glide, dep, etc.
Butcher: Maybe we should do something like dev-v3.
Lots of suggestions for branch names in the chat -- seems that there was some consensus on dev-v3. Butcher will send a message out to the https://lists.cncf.io/g/cncf-kubernetes-helm list.
Assignments
Issue Sherpa: Michelle Noorali
Notes: Nikhil
Moderator: Taylor

March 22, 2018
Announcements
Helm v3 discussions are making progress.
Standup
Adam:
Been working on 1.10 support for Helm. This is now tested.
Worked on vetting LUA models with Butcher.
Fisher:
Pushing through the JFrog remote authentication support for Helm -- that’s done.
Mostly issue triage in addition to that
Taylor:
Was a bit busy this week.
Will take on some more PRs from the PR queue this week -- keep an eye out for reviews especially on mediums and larges.
Going to start trying to look at
Justin
Been pretty busy, so mostly been some issue triaging.
Maybe talk about some regressions that have been creeping up with Helm.
Butcher
Mostly focused on Helm 3
Broke the document out into smaller proposals
Backed out format for chart.yaml -- going to just use the application CRD format so there is less of a learning curve for Helm 3 users
Also been vetting ideas around the LUA extension model -- vetted the loading process, runtime startup, and being able to embed into k8s objects. Cleared the way to say LUA would be the scripting language.
Also simplified the event model this week.
Will keep working at that this week.
Farina
Some more of the Helm 3 stuff
Working on charts -- merged a couple of features upstream wrt better validation of files, linting, etc.
Nikhil
Moved ChartMuseum to dep
Looked at a few PRs and the issue queue.
Plan to work on some of the repo auth pieces this next week.
Discussion
(Josh) Helm repo proposal:
https://github.com/kubernetes-helm/community/pull/18
https://github.com/jdolitsky/community/blob/ebf12b58dec108f85c582369d937100d591f7cdc/helm-v3/006-repositories.md
Methods for authentication -- suggestion to add a helm config for authentication
(Fisher) Have we looked at different authentication models eg. pypi authentication for pip packages?
Another thing to think about is being able to auth against different repos.
This has ramifications on helm search -- you would not be able to search against multiple repos.
Do you do search locally, or do you do search against service providers? This changes the UX in subtle ways since you need to be able to search against certain repos. It would be good to know how people are actually using helm search.
This has similarities with search with pypi -- might be useful to take a look at them too.
During the transition between Helm v2 and v3, it’s really important to be able to search through existing charts and retain backwards compatibility with them.
 (Farina) Do we have concrete requirements from users? Should we go out to users and poll for requirements -- what are the differences in requirements between startups / enterprises.
(Michael Hrivnak -- Welcome!): There is a tremendous amount of value to be had to not requiring a separate service for repositories. Also for enterprise users, they will require content on site. So anything we can do to help with that will be useful.
(Josh) Demo for ChartMuseum multi-tenancy
Assignments
Issue Sherpa: Justin
Notes: Butcher
Moderator: Fisher

March 15, 2018
Announcements
We had a v2.8.2 release go out last week


Standup
Matt Fisher
Helm issue queue stuff
Got a repro for a Helm rollback issue that he’s helping to track down
Nikhil
Spent time looking through the Helm 3 proposal
Helm issue queue stuff
Matt Farina
Working on the Helm 3 User Stories and User Profile docs
Adnan
Reviewing the Helm 3 proposal, user stories
Taylor
Issue sherpa for this week
Question about --reuse-values and its intended behaviour
https://github.com/kubernetes/helm/issues/3685
Justin
Out for a week travelling
Taking a look at the Helm 3 proposal
Matt Butcher
Merged stuff on emeritus core maintainers
Addressing comments on the Helm 3 proposal, moving progress along
Plan for the next week is to help move the Helm 3 stuff along
Kensey
asked https://github.com/kubernetes/charts/issues/4067 for best practices on networking best practices
Discussion
Matt Farina brought up the Helm 3 user stories, saying this is an initial set of user stories to give us direction for whom and for what reasons to help drive Helm 3
Helm 3 Proposal
https://hackmd.io/YvQCGazPRNanSugJV-rbSg
Proposing a client-only model, tiller is gone
Internally, Helm will be restructured around events
Adam mentions notes about Lua
Lua is the easiest to embed
Brian Grant asks if they’ve looked at alternatives like Skylark
Have not looked at Skylark, but they’ve looked at Javascript and found some limitations with regards to constantly catching up with ECMA 6 spec, whereas Lua was designed for embedding
One appendix is to propose using the Lua engine for Helm plugins
Include a `helm push` verb to push helm charts to a repository
Talking to Josh Dolitsky and the ChartMuseum team on defining the spec for `helm push`
Farina notes there’s a lot of proposals also in the Helm issue queue. They’re not noted in the proposal, but they are not being discarded
Next steps: how to break this up into sizeable chunks/proposals
Throw in a milestone?
Move it to issues?
Dump into enhancement proposals/documentation?
Brian asks whether development would happen in the main helm repo
Lazy consensus says yes, we should have a 2.x mainline branch and run Helm 3 development on master
Matt Butcher will publish the Helm 3 proposal in the community repository
Josh added multi-tenancy to ChartMuseum!
Not ready for primetime, but available for preview
Assignments
Issue Sherpa: Matt Fisher
Notes: Nikhil
Moderator: Matt Butcher

March 8, 2018
Announcements
Adam and Butcher have been working on a proposal for Helm 3. Should be ready by next week, when we can discuss it. Taylor has seen a sneak preview.
Taylor gave a brief des
Standup
Taylor:
Going over the Helm 3 proposal stuff
Blog post about the Helm Summit
Will have more time this week for PRs + issues.
Nikhil
Doing some investigation on how OAuth works so he can propose ideas to chartmuseum and other chart repository work
Spent some time responding to Issues
Michelle
Was at the OSLS for the last few days
Working on fetja bot
Need to revisit who should join which mailing list -- maybe create a different ML for the chart maintainers. Added Matt admins to the ML.
Fisher
Last week worked hard on reviewing PRs
Moving forward on discussions on the upgrade bug, and helm init --wait. After discussions with Bryan, Adam, etc came up with solutions that were a lot cleaner that made use of grpc solutions (cool!)
This week, want to pair with Adam on some of the PRs already in the pipeline.
Ronan and Fisher are working on the docs.sh documentation site which is on an old Deis site, trying to move it over to Azure resources. Laptop under someone’s desk => Cloud.
Farina:
Working on getting documentation from the summit together
The user profiles have now been documented at https://github.com/kubernetes-helm/community/blob/master/user-profiles.md
Working on proposing other documentation from the summit.
Poking at the Helm v3 proposal as well.
Adam
Was issue sherpa this week. Thanks to others who jumped in on that.
Also worked on the v3 proposal. Tried to collect all the requirements from the Helm summit and work off of that. Polished that up and put it into documentation based on which we can have a meaningful discussion about. “Too much change could be a bad thing” -- Helm is kicking off, and we might not want to disrupt much. (Something to keep in mind).
Will be pairing to review a couple of hot PRs with Fisher.
Brian
Pairing with Fisher on the Helm wait stuff with the ping using proto health check specification


Discussion
Templating in values.yaml file (see #1978, #2133, #2492, #3252, and https://image.ibb.co/k9Xo07/tpl.png)
Folks would like to go down the road of being able to templatize values.yaml files. Is there a way of going about doing something like that?
Is there a primary use-case for this? The use-cases are documented in the multiple open issues.
Adam: Maybe comes across as an anti-pattern with
One solid use case: Ability to pass the name of a parent chart to a child chart. Does it make sense to solve this use case without templatizing values.yaml?
Other use case: Want to be able to set a value in values.yaml and use that over and over again, later on.
Taylor: Maybe implement this as an alpha feature (as a plugin): if a lot of people use it and find value, maybe move forward then since that may be something that the community wants.
Fisher: Two problems: There are no standard go libs for this. Also in this case people want to render values files before sending to helm. Can we use a plugin to solve some of this (at least initially?)
Adam: It might be possible to get the info in there (eg. parent chart) in a more structured way, rather than by having to templatize values files -- the latter would likely end up with debugging FSM nightmares.
Next steps: read through the issues, to understand the use cases before we can have a deeper discussion next week. Farina will reach out to rheinhardt to get a better idea for use-cases.
Reconnect to tiller on transport error (TODO: issue number -- Farina will follow up)
There are cases where the connection to tiller gets broken. Experiencing this on chart CI and is an outstanding issue.
Suggestion: helm and tiller support a connection via a service that’s created. With CI, it’s good to set up a service and point helm to that service -- you can also use a global portforward. This might be worth looking into for CI. Default helm init might already do this -- something to check.
Another option could be to keep the tunnel open.
Does grpc have an auto-reconnect that we can use?
We see the error in less than a minute on the CI tests so it’s not very long.
E2e tests (see #2846)
Adam will take it over -- has already rebased it. A few more things to clean up. Initial rebase (#3636)
Index.yaml v2 proposal (#3557)
Farina: Can we get more eyeballs on this?
This goes hand-in-hand with the scaling discussion for chart repositories. Has implications on the growing size of the community charts repositories.
We will try and cover this a bit more during the next meeting.
ChartMuseum OAuth discussion: https://github.com/kubernetes-helm/chartmuseum/issues/59
Assignments
Issue Sherpa: Taylor with Michelle as assistant to the regional sherpa! :)
Notes: Fisher
Moderator: Adam

March 1, 2018
Announcements
Helm Summit was awesome last week; nice job everyone!
Thanks to Fisher for closing a massive amount of PRs
Standup
Fisher:
Post summit stuff
Reviewing blog post with Taylor
Follow up with Karen (Microsoft) on post-summit feedback email
Looked at an merged a whole bunch of PRs (Thanks Fisher!)
This week: More PRs + Follow up on Helm 3 discussions, and next steps from notes from the Summit
Taylor:
Distilling requirements for Helm 3 from unconference discussions
Should discuss these requirements in a follow-up meeting
Nikhil:
Issue sherpa’ing this week
Post Helm 3 discussions with josh Dolitsky on ChartMuseum multi-tenancy stuff, feedback
Will have a follow-up discussion with Josh on auth in ChartMuseum
Adam:
Did a few PR reviews
Fixed a few bugs with Helm plugins
Updated all his plugins
Bucher and Adam got together, writing and sketching a proposal for Helm 3 with a few decision trees, which should be good for discussions
Butcher’s away in Texas this week, so it might not be finished until next week
Justin:
Popped in late, talking about triaging issues this week
Proposed an issue template on kubernetes-helm-core for Helm
Largely positive feedback on the proposal, so Taylor asks about proposing a PR to it
Asked about Brian Grant’s proposal, which
Discussion
Fisher: talk about https://github.com/kubernetes/helm/pull/3437
Adam explained how upgrading as if it was a fresh install won’t work due to the install being in an unknown state
Fisher agreed
Taylor proposed perhaps this would be a good feature flag instead? That way users can opt into destructive behaviour
Adam and Fisher both agreed that would be a good proposal
Adam will bring some context back into the PR so we can hopefully steer the conversation in that direction
Nikhil: When sherpa’ing, how can we determine which features should be frozen from fejta-bot?
Adam and Taylor provided context and info on asking questions about older issues on helm-dev
Pairing sessions are great for this!
Nikhil agreed
Assignments
Issue Sherpa: Adam
Notes: Nikhil
Moderator: Taylor?
February 15, 2018
Announcements
Helm summit next week! http://helmsummitpdx-feb2018.splashthat.com Get your tickets if you don’t already have them!


Standup
Justin
Mostly issues and PRs
Is excited(?) to be the AV guy at the summit at 7:30. :) We appreciate the contributions. There will be coffee!
Matt Butcher
Helm summit things
Collecting some feedback on what are currently the most requested features for Helm -- to try and steer the Helm 3 discussion that way.
What is the long term structure / governance model for Helm? Is this something we want to discuss at the summit?
Open question: Is it better to lay down (discuss) suggested requirements, rather than dictate solutions for Helm 3?
Matt Fisher
Majority of the work was helping out with the Dev summit
Working with Ronan? To figure out logistics regarding the summit.
Michelle
Worked on reviewing PR (https://github.com/kubernetes/helm/pull/3206) -- is now up for review
Updated the readme with the latest mailing lists from CNCF. Focused on making sure they’re set up correctly this week as well. There are multiple lists -- users, reviewers, power users, etc. Please join any relevant mailing lists.
https://lists.cncf.io/g/cncf-kubernetes-helm
https://lists.cncf.io/g/cncf-kubernetes-helm-power-users
https://lists.cncf.io/g/cncf-kubernetes-helm-security
https://lists.cncf.io/g/cncf-kubernetes-helm-maintainers


Also working on speaker stuff for Helm summit


Taylor
Working on Helm summit things -- big shout out to all the folks working on Helm summit items (Michelle, Adam, Fisher, everyone else)
Nikhil
Working on updating SUSE’s version of helm along with ChartMuseum
Working on his talk for the summit
(Nikhil: Thanks Fisher for noting this down while I was speaking!)
Adam
More summit stuff
Focused on summit talks and speaker work items
Brian Hardock
Just listening in this week
Discussion
There are some bugs in 2.8
Helm list maximum client size exceeded (Justin’s case -- actually exceeding the 20 MB limit) 2.7 -> 2.8 bumped grpc, which made the client and server use the same size. Balboa has a new PR which updated the server client-receive size as well (apparently there are 3 settings that need to be updated). If this does fix it, does this warrant a 2.8.2, and if so -- do we wait until the summit to cut a 2.8.2?
Taylor / Butcher / Nikhil: Doesn’t seem like too many folks are affected by this - so probably prudent to wait until after summit.
Don’t need a dev call next week since we’ll be at the summit.
Michelle will send out an announcement on the ML and _all the places_, that we are at the summit next week, so things might be a bit slow.
Taylor: Maybe we should also put a message in the dev channel that we are at the summit, in case some high-pri issues do come up. (“Emergency mode”)
Assignments (NOTE - assignments are for March 1st -- the week _after_ the summit)
Issue sherpa: Nikhil
Notes: Michelle
Moderator: Adam

February 8, 2018
Announcements
Helm summit it comes soon. Tell your friends. http://helmsummitpdx-feb2018.splashthat.com
Standup
Justin
Issue triaging
Tried to generate the docs site working on issues and it failed. We have an issue
Matt Farina
Prepping for Helm summit talks
Fixed charts CI woes (ran out of resources due to memory leak)
PR to fix GRPC bug from update
Reviewing 2.8.1 issues/PRs
Being involved with KEP and CRD discussions for application type.
Matt Butcher
Looking into images to docker hub and other places for put binaries for better metrics that are public
Looking to clear some schedule time for helm development
Matt Fisher
Triaged things for 2.8.1
Dove into bug of multiple releases being marked as deployed
Going to pair with Butcher on keynote which isn’t done yet
Going to work to get 2.8.1 out
Michelle
Reviewed on auth PR (by jfrog)
Will be around to review more things
CNCF requested mailing lists are setup
Taylor
Helm summit- all the things!
Nikhil
Issue triage
Worked on issue with chartmuseum on helm and updates
Worked on his helm summit talk
Adam
Behind on everything so not too much to report. Took time off
Discussion
Discussed https://github.com/kubernetes/helm/issues/3242
2.8.1 burn down
There are two PRs left that need reviewing
Mailing lists
Didn’t get many people biting on power users list but will announce at the summit
Helm site
CI is still deis and needs to move elsewhere as that is going away in March
Assignments
Issue sherpa: Matt Butcher
Notes: Nikhil
Moderator: Matt Fisher

February 1, 2018
Announcements
<No announcements>
Standup
Justin
Mostly looking at issues
Looking at reuse values bug
Trying to help people having one-off issues.
Matt Farina
Upgrade bug for 2.8.0 which might hit people. Might need to issue a fix and cut a 2.8.1 release for issue 3382
Odd testing scenarios and poking at CI things.
Taylor
Was mostly doing Helm Summit stuff
Might not have too much time for issue sherpa given the Summit tasks
Nikhil
Looking at issues and PRs
Progress towards Helm Summit talk
Discussion
Kubernetes is going to shift away from the term “Master” (eg. drupal went with primary and replica, etc.) Propose to use something like API Server / Control Plane. There are places in docs where we use the term “Master” -- this needs to be updated.
Need a few more eyes on issue 3382 to figure out why upgrades are failing. Seems to be an issue with patch metadata.
We fixed a bug where we forgot to bump the version of the helm cli internally. And a couple of other bugs. Do we fix 3382 and cut a 2.8.1, or do we cut 2.8.1 without a fix for this since there are other fixes that we want to get out?
Was specifically asked to be noted that we finished 10 minutes early without MS folks around ;)
Assignments
Issue sherpa: Justin
Notes: Matt Farina
Moderator: Taylor
January 25, 2018
Announcements
Helm Summit registration is open <insert link>
Standup
Matt Fisher
Paired with Farina on release cutting for 2.8.0 off RC1
Issues are coming in related to Kubernetes 1.9.0 libraries
Issue triaging this week, not around next week as much
Aggressively reviewing PRs
Adnan
Interesting things going on in the chart repo
Working on Kube Apps related things
New dashboard rewrite
Taylor
Helm Summit
Working on CRD talk for Helm Summit. Will ping asking for feedback
Matt Farina
Followed along on the release process which is nicely documented
Issue open about signing Helm releases. Looking at that more closely and how other projects to do this to respond.
Poking at bug on helm 2.8.0 that has to do with Helm upgrades.
On the charts side, usual stuff
Also documenting more processes on the charts repo
Also documenting public chart best practices
Helps give people clearer idea on expectations
Matt Butcher


Brian Hardock
Listening today
Nikhil
Issue triaging this week
In Germany this week talking about Helm at SUSE
Preparing for the Helm Summit
Michelle Noorali
Helm summit schedule is live, schedule is up
44% of the available tickets have been sold
Discussion
Onboarding documents
Put in Kubernetes-Helm org in the community repo
Transparent governance
Helm Summit
Core maintainers who will be there
Matt Butcher, Matt Fisher, Matt Farina, Brian Hardock, Nikhil
There will be a Q&A with a core maintainer
Matt Fisher volunteered for unconference organization
Miguel and Scott Rigby
(Taylor) How we publicize RCs that may have prevent bugs
Let’s collect a group of Helm power users to test RCs
Problem with testing charts in RCs
CNCF is switching over to groups.io and may be able to set up some mailing lists for us
Michelle to follow up
CNCF has already switched (https://lists.cncf.io/g/main) - thnx for the link Farina
Assignments
Issue sherpa: Matt Fisher
Notes: Nikhil
Moderator:
January 18, 2018
Announcements
Helm Dev Summit CFP acceptance letters have been sent out, just finalizing the details and schedule
Tickets are set at $75, anyone who RSVP’d will get an email with the link
Standup
Taylor
Working primarily on Helm Summit stuff
Finalizing some of those details this coming week
Has a lightning talk; would like to bug others for feedback
Matt Farina
Found out that .Capabilities.GitVersion is missing for testing purposes
Currentl workflow is string comparison so x.9 > x.10
Nikhil
Shadowed the issue queue w/ Matt Fisher
Reminded that he will be out next week, putting information together for SUSE folks
Wanted to follow up with Butcher on the onboarding documentation
Matt Butcher
Wanted to announce that Matt Farina and Nikhil are both core maintainers
Been on the customer front talking about helm
Nailed down a bug with `helm list` in related to rolling back and recovering lost errors
Looking to hire Helm developers @ Microsoft
Adam
Mostly on vacation last week and the following week
Pairing with Matt Butcher on the `helm list` bug
Helm 2.8 release stuff
Matt Fisher
CFP submission scheduling
Discussion
Come to a final decision on how we are going to use test-infra and any concerns we have with it
Farina gave a demo on https://prow.k8s.io/plugin-help.html
Butcher: can we version pin plugins/opt out of new features e.g. new labels introduced in the label plugin?
Farina: policy on how we roll this out is discussing right now in SIG test infra
Feedback: adding too much heavy-handed test automation is not required for smaller projects like Helm; it does not need a whole lot of tooling because the process is intentionally simple
Farina is hoping more people to come to SIG contribex to provide this feedback
Best option is to look at the plugin repo to see what we can opt in (and out) of
Building for other architectures
Opinions on releasing a separate docker image for Helm vs Tiller?
Going forward with splitting into separate docker images
Looking forward to supporting alternate architectures
Publish these images to Dockerhub? Conversations leaned towards “yes”
Helm v2.8 release
Planning to cut the release next week, Fisher will take the lead. Farina and Taylor are interested in pairing to understand the release process
Assignments
Issue Sherpa: Nikhil
Meeting Lead: Matt Butcher
Meeting Notes: Michelle (voluntold)

January 11, 2018 [Recording]
Announcements
Helm 2.8.0-rc1 is released
Helm 2.8.0 is being targeted for release next week
Standup
Matt Fisher
Helm Summit CFP reviewing
Justin
Reviewed some PRs
Taylor
Reviewing CFPs, helm summit stuff
Nik
PR reviews
Worked with Matt Butcher to put together some material to talk about Helm at SUSE
This week wants to work on learning how to issue czar
Michelle
Helm Summit CFP review and planning
Same thing for this week
Reminder: there is a core maintainer meeting the night before the Helm Summit
Matt Farina
Helped merge the feature for the supported Kubernetes versions for a chart
Adam
Helm Summit CFP reviewing
This week is more Helm Summit stuff
Getting ready for 2.8 release
Discussion
RBAC best practices changes
Do we create a service account for each service by default (in Tiller)?
A service account per service is a pretty good idea
There is a debate about whether or not we should have Tiller do this.
The resolution is that we go forward with adding service account creation in the charts themselves and not tiller
Feedback requested: Common labels (mattfarina)
Common labels allow for some really good possibilities for visualization
K8s tooling (/lgtm, /approve and owners file changes) (mattfarina)
Kubernetes is wanting everyone to use the test-infra tools. Do we want to use them?
There are some useful testing functions that the tools enable, like tests right before a merge
If we have issues with having these enforced, we need to let people know at https://github.com/kubernetes/test-infra
CRD stuff (Justin)
Moved to next week. This is about changing the validation piece so CRDs function properly
We are ok if someone wants to refactor the validation
Assignments
Issue Czar: Matt Fisher with Nik
Meeting Notes: Michelle
Meeting Lead: Taylor
January 4, 2018
Announcements
Helm Summit CFP is officially closed
Standup
Taylor
Mostly working on clearing out notifications on github
Reviewed adam’s PR for kube 1.9
Reviewing CFPs for the Helm Summit
Matt Butcher
2.8 release planning is priority
Matt Fisher
See Taylor’s standup
Slicknik
Wants to sync up with butcher to ramp up SUSE on Helm
Adam
See Taylor’s standup
Discussion
Paul (withnale) talked about https://github.com/kubernetes/helm/pull/3243
Butcher and Taylor mention it would be good to get Justin Scott to take a look at this
Helm 2.8
Nothing else blocking except for 1.9 support
We should continue to cut an RC like the last few releases
Assignments
Issue sherpa: Matt Butcher
Note Taker: Taylor
Meeting Lead: Adam Reese